Automated stratification of trauma injury severity across multiple body regions using multi-modal, multi-class machine learning models

OBJECTIVE

The timely stratification of trauma injury severity can enhance the quality of trauma care but it requires intense manual annotation from certified trauma coders. The objective of this study is to develop machine learning models for the stratification of trauma injury severity across various body regions using clinical text and structured electronic health records (EHRs) data.

MATERIALS AND METHODS

Our study utilized clinical documents and structured EHR variables linked with the trauma registry data to create 2 machine learning models with different approaches to representing text. The first one fuses concept unique identifiers (CUIs) extracted from free text with structured EHR variables, while the second one integrates free text with structured EHR variables. Temporal validation was undertaken to ensure the models' temporal generalizability. Additionally, analyses to assess the variable importance were conducted.

RESULTS

Both models demonstrated impressive performance in categorizing leg injuries, achieving high accuracy with macro-F1 scores of over 0.8. Additionally, they showed considerable accuracy, with macro-F1 scores exceeding or near 0.7, in assessing injuries in the areas of the chest and head. We showed in our variable importance analysis that the most important features in the model have strong face validity in determining clinically relevant trauma injuries.

DISCUSSION

The CUI-based model achieves comparable performance, if not higher, compared to the free-text-based model, with reduced complexity. Furthermore, integrating structured EHR data improves performance, particularly when the text modalities are insufficiently indicative.

CONCLUSIONS

Our multi-modal, multiclass models can provide accurate stratification of trauma injury severity and clinically relevant interpretations.

LCD Benchmark: Long Clinical Document Benchmark on Mortality Prediction

Natural Language Processing (NLP) is a study of automated processing of text data. Application of NLP in the clinical domain is important due to the rich unstructured information implanted in clinical documents, which often remains inaccessible in structured data. Empowered by the recent advance of language models (LMs), there is a growing interest in their application within the clinical domain. When applying NLP methods to a certain domain, the role of benchmark datasets are crucial as benchmark datasets not only guide the selection of best-performing models but also enable assessing of the reliability of the generated outputs. Despite the recent availability of LMs capable of longer context, benchmark datasets targeting long clinical document classification tasks are absent. To address this issue, we propose LCD benchmark, a benchmark for the task of predicting 30-day out-of-hospital mortality using discharge notes of MIMIC-IV and statewide death data. Our notes have a median word count of 1687 and an interquartile range of 1308 to 2169. We evaluated this benchmark dataset using baseline models, from bag-of-words and CNN to Hierarchical Transformer and an open-source instruction-tuned large language model. Additionally, we provide a comprehensive analysis of the model outputs, including manual review and visualization of model weights, to offer insights into their predictive capabilities and limitations. We expect LCD benchmarks to become a resource for the development of advanced supervised models, prompting methods, or the foundation models themselves, tailored for clinical text. The benchmark dataset is available at https://github.com/Machine-Learning-for-Medical-Language/long-clinical-doc.

Achievement of Target Gain Larger than Unity in an Inertial Fusion Experiment

On December 5, 2022, an indirect drive fusion implosion on the National Ignition Facility (NIF) achieved a target gain G_{target} of 1.5. This is the first laboratory demonstration of exceeding "scientific breakeven" (or G_{target}>1) where 2.05 MJ of 351 nm laser light produced 3.1 MJ of total fusion yield, a result which significantly exceeds the Lawson criterion for fusion ignition as reported in a previous NIF implosion [H. Abu-Shawareb et al. (Indirect Drive ICF Collaboration), Phys. Rev. Lett. 129, 075001 (2022)PRLTAO0031-900710.1103/PhysRevLett.129.075001]. This achievement is the culmination of more than five decades of research and gives proof that laboratory fusion, based on fundamental physics principles, is possible. This Letter reports on the target, laser, design, and experimental advancements that led to this result.

A fluid biomarker reveals loss of TDP-43 splicing repression in presymptomatic ALS-FTD

Although loss of TAR DNA-binding protein 43 kDa (TDP-43) splicing repression is well documented in postmortem tissues of amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD), whether this abnormality occurs during early-stage disease remains unresolved. Cryptic exon inclusion reflects loss of function of TDP-43, and thus detection of proteins containing cryptic exon-encoded neoepitopes in cerebrospinal fluid (CSF) or blood could reveal the earliest stages of TDP-43 dysregulation in patients. Here we use a newly characterized monoclonal antibody specific to a TDP-43-dependent cryptic epitope (encoded by the cryptic exon found in HDGFL2) to show that loss of TDP-43 splicing repression occurs in ALS-FTD, including in presymptomatic C9orf72 mutation carriers. Cryptic hepatoma-derived growth factor-like protein 2 (HDGFL2) accumulates in CSF at significantly higher levels in familial ALS-FTD and sporadic ALS compared with controls and is elevated earlier than neurofilament light and phosphorylated neurofilament heavy chain protein levels in familial disease. Cryptic HDGFL2 can also be detected in blood of individuals with ALS-FTD, including in presymptomatic C9orf72 mutation carriers, and accumulates at levels highly correlated with those in CSF. Our findings indicate that loss of TDP-43 cryptic splicing repression occurs early in disease progression, even presymptomatically, and that detection of the HDGFL2 cryptic neoepitope serves as a potential diagnostic biomarker for ALS, which should facilitate patient recruitment and measurement of target engagement in clinical trials.

Standardizing Multi-site Clinical Note Titles to LOINC Document Ontology: A Transformer-based Approach

The types of clinical notes in electronic health records (EHRs) are diverse and it would be great to standardize them to ensure unified data retrieval, exchange, and integration. The LOINC Document Ontology (DO) is a subset of LOINC that is created specifically for naming and describing clinical documents. Despite the efforts of promoting and improving this ontology, how to efficiently deploy it in real-world clinical settings has yet to be explored. In this study we evaluated the utility of LOINC DO by mapping clinical note titles collected from five institutions to the LOINC DO and classifying the mapping into three classes based on semantic similarity between note titles and LOINC DO codes. Additionally, we developed a standardization pipeline that automatically maps clinical note titles from multiple sites to suitable LOINC DO codes, without accessing the content of clinical notes. The pipeline can be initialized with different large language models, and we compared the performances between them. The results showed that our automated pipeline achieved an accuracy of 0.90. By comparing the manual and automated mapping results, we analyzed the coverage of LOINC DO in describing multi-site clinical note titles and summarized the potential scope for extension.

An End-to-End Natural Language Processing System for Automatically Extracting Radiation Therapy Events From Clinical Texts

PURPOSE

Real-world evidence for radiation therapy (RT) is limited because it is often documented only in the clinical narrative. We developed a natural language processing system for automated extraction of detailed RT events from text to support clinical phenotyping.

METHODS AND MATERIALS

A multi-institutional data set of 96 clinician notes, 129 North American Association of Central Cancer Registries cancer abstracts, and 270 RT prescriptions from HemOnc.org was used and divided into train, development, and test sets. Documents were annotated for RT events and associated properties: dose, fraction frequency, fraction number, date, treatment site, and boost. Named entity recognition models for properties were developed by fine-tuning BioClinicalBERT and RoBERTa transformer models. A multiclass RoBERTa-based relation extraction model was developed to link each dose mention with each property in the same event. Models were combined with symbolic rules to create a hybrid end-to-end pipeline for comprehensive RT event extraction.

RESULTS

Named entity recognition models were evaluated on the held-out test set with F1 results of 0.96, 0.88, 0.94, 0.88, 0.67, and 0.94 for dose, fraction frequency, fraction number, date, treatment site, and boost, respectively. The relation model achieved an average F1 of 0.86 when the input was gold-labeled entities. The end-to-end system F1 result was 0.81. The end-to-end system performed best on North American Association of Central Cancer Registries abstracts (average F1 0.90), which are mostly copy-paste content from clinician notes.

CONCLUSIONS

We developed methods and a hybrid end-to-end system for RT event extraction, which is the first natural language processing system for this task. This system provides proof-of-concept for real-world RT data collection for research and is promising for the potential of natural language processing methods to support clinical care.

The temporal burden of preparing catheters for re-use in adults with spinal cord injury: a cross-sectional study

STUDY DESIGN

Cross-sectional OBJECTIVES: Neurogenic lower urinary tract dysfunction is common among people with spinal cord injury (SCI). Although single-use clean intermittent catheterization is recommended to facilitate routine bladder emptying, catheter re-use is common. Barriers associated with the preparation (i.e., cleaning) of catheters for re-use are unknown. This study examined barriers to catheter re-use in adult individuals with SCI by assessing (1) the time needed to clean a catheter, and (2) the perceived difficulty of the catheter cleaning routine.

SETTING

Laboratory METHODS: Twenty individuals with chronic SCI ( ≥ 1 year since injury; Group 1 = 10 with tetraplegia; Group 2 = 10 with paraplegia) completed the study. Using a standardized cleaning procedure (i.e., Milton method), catheter cleaning was timed for each participant. Perceived difficulty was assessed using a 5-point Likert scale. Functional impairment was assessed with the Upper Extremity Motor Score (UEMS).

RESULTS

Significant between-group differences were observed for total cleaning time (Group 1 = 1584.1 ± 179.8 s; Group 2 = 1321.0 ± 93.8 s, p = 0.004) and perceived difficulty [Group 1 = 2.6 (2, 3); Group 2 = 2 (1.7, 2.3), p = 0.028]. Total cleaning time was significantly correlated with UEMS (ρ = -0.709, p ≤ 0.001) and perceived difficulty (ρ = 0.468, p = 0.037). UEMS emerged as an independent predictor of total cleaning time (R² = 0.745, β = -0.833, p ≤ 0.001).

CONCLUSIONS

Preparing catheters for re-use is time-intensive and difficult for people with higher SCI level, severity and more pronounced upper limb motor impairment, which was independently associated with total cleaning time. Performing this routine on a consistent basis would require a substantial time commitment and would have a profoundly negative impact on overall quality of life.

Chemical synthesis of amphiphilic glycoconjugates: Access to amino, fluorinated and sulfhydryl oleyl glucosides

Amphiphilic glycoconjugates offer an important prospect for development as chemical biology tools and biosurfactants. The chemical synthesis of such materials is required to expedite such prospect, compounded by the example of oleyl glycosides. Herein, we report a mild and reliable glycosylation method to access oleyl glucosides, glycosidating oleyl alcohol with α-trichloroacetimidate donors. We demonstrate capability for this methodology, extending it to synthesise the first examples of pyranose-component fluorination and sulfhydryl modifications within glucosides and glucosamines of oleyl alcohol. These compounds provide an exciting series of tools to explore processes and materials that utilise oleyl glycosides, including as probes for glycosphingolipid metabolism.

End-to-end clinical temporal information extraction with multi-head attention

Understanding temporal relationships in text from electronic health records can be valuable for many important downstream clinical applications. Since Clinical TempEval 2017, there has been little work on end-to-end systems for temporal relation extraction, with most work focused on the setting where gold standard events and time expressions are given. In this work, we make use of a novel multi-headed attention mechanism on top of a pre-trained transformer encoder to allow the learning process to attend to multiple aspects of the contextualized embeddings. Our system achieves state of the art results on the THYME corpus by a wide margin, in both the in-domain and cross-domain settings.

Multi-Task Training with In-Domain Language Models for Diagnostic Reasoning

Generative artificial intelligence (AI) is a promising direction for augmenting clinical diagnostic decision support and reducing diagnostic errors, a leading contributor to medical errors. To further the development of clinical AI systems, the Diagnostic Reasoning Benchmark (DR.BENCH) was introduced as a comprehensive generative AI framework, comprised of six tasks representing key components in clinical reasoning. We present a comparative analysis of in-domain versus out-of-domain language models as well as multi-task versus single task training with a focus on the problem summarization task in DR.BENCH (Gao et al., 2023). We demonstrate that a multi-task, clinically-trained language model outperforms its general domain counterpart by a large margin, establishing a new state-of-the-art performance, with a ROUGE-L score of 28.55. This research underscores the value of domain-specific training for optimizing clinical diagnostic reasoning tasks.

Overview of the Problem List Summarization (ProbSum) 2023 Shared Task on Summarizing Patients' Active Diagnoses and Problems from Electronic Health Record Progress Notes

The BioNLP Workshop 2023 initiated the launch of a shared task on Problem List Summarization (ProbSum) in January 2023. The aim of this shared task is to attract future research efforts in building NLP models for real-world diagnostic decision support applications, where a system generating relevant and accurate diagnoses will augment the healthcare providers' decision-making process and improve the quality of care for patients. The goal for participants is to develop models that generated a list of diagnoses and problems using input from the daily care notes collected from the hospitalization of critically ill patients. Eight teams submitted their final systems to the shared task leaderboard. In this paper, we describe the tasks, datasets, evaluation metrics, and baseline systems. Additionally, the techniques and results of the evaluation of the different approaches tried by the participating teams are summarized.

Representing and utilizing clinical textual data for real world studies: An OHDSI approach

Clinical documentation in electronic health records contains crucial narratives and details about patients and their care. Natural language processing (NLP) can unlock the information conveyed in clinical notes and reports, and thus plays a critical role in real-world studies. The NLP Working Group at the Observational Health Data Sciences and Informatics (OHDSI) consortium was established to develop methods and tools to promote the use of textual data and NLP in real-world observational studies. In this paper, we describe a framework for representing and utilizing textual data in real-world evidence generation, including representations of information from clinical text in the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM), the workflow and tools that were developed to extract, transform and load (ETL) data from clinical notes into tables in OMOP CDM, as well as current applications and specific use cases of the proposed OHDSI NLP solution at large consortia and individual institutions with English textual data. Challenges faced and lessons learned during the process are also discussed to provide valuable insights for researchers who are planning to implement NLP solutions in real-world studies.

Progress Note Understanding - Assessment and Plan Reasoning: Overview of the 2022 N2C2 Track 3 shared task

Daily progress notes are a common note type in the electronic health record (EHR) where healthcare providers document the patient's daily progress and treatment plans. The EHR is designed to document all the care provided to patients, but it also enables note bloat with extraneous information that distracts from the diagnoses and treatment plans. Applications of natural language processing (NLP) in the EHR is a growing field with the majority of methods in information extraction. Few tasks use NLP methods for downstream diagnostic decision support. We introduced the 2022 National NLP Clinical Challenge (N2C2) Track 3: Progress Note Understanding - Assessment and Plan Reasoning as one step towards a new suite of tasks. The Assessment and Plan Reasoning task focuses on the most critical components of progress notes, Assessment and Plan subsections where health problems and diagnoses are contained. The goal of the task was to develop and evaluate NLP systems that automatically predict causal relations between the overall status of the patient contained in the Assessment section and its relation to each component of the Plan section which contains the diagnoses and treatment plans. The goal of the task was to identify and prioritize diagnoses as the first steps in diagnostic decision support to find the most relevant information in long documents like daily progress notes. We present the results of the 2022 N2C2 Track 3 and provide a description of the data, evaluation, participation and system performance.

Automatic Extraction of Medication Mentions from Tweets-Overview of the BioCreative VII Shared Task 3 Competition

This study presents the outcomes of the shared task competition BioCreative VII (Task 3) focusing on the extraction of medication names from a Twitter user's publicly available tweets (the user's 'timeline'). In general, detecting health-related tweets is notoriously challenging for natural language processing tools. The main challenge, aside from the informality of the language used, is that people tweet about any and all topics, and most of their tweets are not related to health. Thus, finding those tweets in a user's timeline that mention specific health-related concepts such as medications requires addressing extreme imbalance. Task 3 called for detecting tweets in a user's timeline that mentions a medication name and, for each detected mention, extracting its span. The organizers made available a corpus consisting of 182 049 tweets publicly posted by 212 Twitter users with all medication mentions manually annotated. The corpus exhibits the natural distribution of positive tweets, with only 442 tweets (0.2%) mentioning a medication. This task was an opportunity for participants to evaluate methods that are robust to class imbalance beyond the simple lexical match. A total of 65 teams registered, and 16 teams submitted a system run. This study summarizes the corpus created by the organizers and the approaches taken by the participating teams for this challenge. The corpus is freely available at https://biocreative.bioinformatics.udel.edu/tasks/biocreative-vii/track-3/. The methods and the results of the competing systems are analyzed with a focus on the approaches taken for learning from class-imbalanced data.

DR.BENCH: Diagnostic Reasoning Benchmark for Clinical Natural Language Processing

The meaningful use of electronic health records (EHR) continues to progress in the digital era with clinical decision support systems augmented by artificial intelligence. A priority in improving provider experience is to overcome information overload and reduce the cognitive burden so fewer medical errors and cognitive biases are introduced during patient care. One major type of medical error is diagnostic error due to systematic or predictable errors in judgement that rely on heuristics. The potential for clinical natural language processing (cNLP) to model diagnostic reasoning in humans with forward reasoning from data to diagnosis and potentially reduce cognitive burden and medical error has not been investigated. Existing tasks to advance the science in cNLP have largely focused on information extraction and named entity recognition through classification tasks. We introduce a novel suite of tasks coined as Diagnostic Reasoning Benchmarks, Dr.Bench, as a new benchmark for developing and evaluating cNLP models with clinical diagnostic reasoning ability. The suite includes six tasks from ten publicly available datasets addressing clinical text understanding, medical knowledge reasoning, and diagnosis generation. DR.BENCH is the first clinical suite of tasks designed to be a natural language generation framework to evaluate pre-trained language models for diagnostic reasoning. The goal of DR. BENCH is to advance the science in cNLP to support downstream applications in computerized diagnostic decision support and improve the efficiency and accuracy of healthcare providers during patient care. We fine-tune and evaluate the state-of-the-art generative models on DR.BENCH. Experiments show that with domain adaptation pre-training on medical knowledge, the model demonstrated opportunities for improvement when evaluated in DR. BENCH. We share DR. BENCH as a publicly available GitLab repository with a systematic approach to load and evaluate models for the cNLP community. We also discuss the carbon footprint produced during the experiments and encourage future work on DR.BENCH to report the carbon footprint.

Target Ultra and Nano coils in the endovascular treatment of small intracranial aneurysms (ULTRA Registry)

OBJECTIVE

The ULTRA Registry is a national multicenter prospective study designed to assess aneurysm occlusion rates and safety profiles of the Target Ultra and Nano coils in the treatment of small intracranial aneurysms (IAs).

METHODS

Patients with small (≤ 5 mm) ruptured and unruptured IAs were treated exclusively with Target Ultra and Nano coils. The primary endpoints were the initial rate of complete or near-complete aneurysm occlusion, aneurysm recurrence, and need for retreatment. Secondary endpoints were device- and procedure-related adverse events, hemorrhage from the coiled aneurysm at any time during follow-up, and clinical outcomes.

RESULTS

The ULTRA Registry included 100 patients with a mean ± SD age of 56 ± 11.6 years, of whom 75 were women and 48 presented after aneurysm rupture. The mean aneurysm size was (3.5 ± 0.9) × (2.8 ± 0.9) × (3.0 ± 1.0) mm, and the mean packing density was 34.4% ± 16.7%. Posttreatment complete or near-complete occlusion reported by an independent imaging core laboratory was seen in 92% of patients at baseline and in 87%, 87%, and 83% of patients at first, second, and final follow-up, respectively. At first, second, and final follow-up, 10%, 11%, and 15%, respectively, of patients were deemed to require retreatment. There were three procedural-related ischemic strokes and one intracranial hemorrhage from wire perforation of a parent artery not involved by the aneurysm. There were no coil-related adverse events, including no intraoperative aneurysm ruptures and no known aneurysm ruptures after coiling.

CONCLUSIONS

This assessment of aneurysm occlusion rates and safety profiles in ULTRA Registry study participants demonstrates excellent safety and efficacy profiles for Target Ultra and Nano coils in the treatment of small IAs.

Psychology and AI at a Crossroads: How Might Complex Systems Explain Themselves?

A challenge in building useful artificial intelligence (AI) systems is that people need to understand how they work in order to achieve appropriate trust and reliance. This has become a topic of considerable interest, manifested as a surge of research on Explainable AI (XAI). Much of the research assumes a model in which the AI automatically generates an explanation and presents it to the user, whose understanding of the explanation leads to better performance. Psychological research on explanatory reasoning shows that this is a limited model. The design of XAI systems must be fully informed by a model of cognition and a model of pedagogy, based on empirical evidence of what happens when people try to explain complex systems to other people and what happens as people try to reason out how a complex system works. In this article we discuss how and why C. S. Peirce's notion of abduction is a best model for XAI. Peirce's notion of abduction as an exploratory activity can be regarded as supported by virtue of its concordance with models of expert reasoning that have been developed by modern applied cognitive psychologists.

The SOD1-mediated ALS phenotype shows a decoupling between age of symptom onset and disease duration

Superoxide dismutase (SOD1) gene variants may cause amyotrophic lateral sclerosis, some of which are associated with a distinct phenotype. Most studies assess limited variants or sample sizes. In this international, retrospective observational study, we compare phenotypic and demographic characteristics between people with SOD1-ALS and people with ALS and no recorded SOD1 variant. We investigate which variants are associated with age at symptom onset and time from onset to death or censoring using Cox proportional-hazards regression. The SOD1-ALS dataset reports age of onset for 1122 and disease duration for 883 people; the comparator population includes 10,214 and 9010 people respectively. Eight variants are associated with younger age of onset and distinct survival trajectories; a further eight associated with younger onset only and one with distinct survival only. Here we show that onset and survival are decoupled in SOD1-ALS. Future research should characterise rarer variants and molecular mechanisms causing the observed variability.

Technique for Reconstruction of the Proximal Tibiofibular Joint Using Biceps Femoris and Iliotibial Band Autograft With Suture Tape Augmentation

Background:

Proximal tibiofibular joint instability is an uncommon condition that can occur in acute traumatic dislocations, chronic or recurrent dislocations, and atraumatic dislocations. A variety of procedures have been described for treatment of this condition including fibular head resection, arthrodesis, suture button stabilization, and soft tissue reconstruction.

Indications:

The initial management consists of conservative measures such as activity modification, use of a supportive strap, and physical therapy. When nonoperative management fails, surgical management is indicated.

Technique Description:

To optimize surgical technique and patient outcomes, a previously described reconstruction technique was modified with reinforcement of suture tape augmentation and decreased invasiveness.

Results:

This technique has several advantages including a smaller incision with shorter iliotibial band (ITB) and biceps femoris tendon (BFT) grafts and safer posterior to anterior drilling as opposed to anterior to posterior directed drilling.

Discussion/Conclusion:

The authors believe there is increased stability with this reconstruction technique due to the added suture tape augmentation in addition to the biologic healing potential provided by the ITB and BFT autografts. The author(s) attests that consent has been obtained from any patient(s) appearing in this publication. If the individual may be identifiable, the author(s) has included a statement of release or other written form of approval from the patient(s) with this submission for publication.

Summarizing Patients' Problems from Hospital Progress Notes Using Pre-trained Sequence-to-Sequence Models

Automatically summarizing patients' main problems from daily progress notes using natural language processing methods helps to battle against information and cognitive overload in hospital settings and potentially assists providers with computerized diagnostic decision support. Problem list summarization requires a model to understand, abstract, and generate clinical documentation. In this work, we propose a new NLP task that aims to generate a list of problems in a patient's daily care plan using input from the provider's progress notes during hospitalization. We investigate the performance of T5 and BART, two state-of-the-art seq2seq transformer architectures, in solving this problem. We provide a corpus built on top of progress notes from publicly available electronic health record progress notes in the Medical Information Mart for Intensive Care (MIMIC)-III. T5 and BART are trained on general domain text, and we experiment with a data augmentation method and a domain adaptation pre-training method to increase exposure to medical vocabulary and knowledge. Evaluation methods include ROUGE, BERTScore, cosine similarity on sentence embedding, and F-score on medical concepts. Results show that T5 with domain adaptive pre-training achieves significant performance gains compared to a rule-based system and general domain pre-trained language models, indicating a promising direction for tackling the problem summarization task.

Lawson Criterion for Ignition Exceeded in an Inertial Fusion Experiment

For more than half a century, researchers around the world have been engaged in attempts to achieve fusion ignition as a proof of principle of various fusion concepts. Following the Lawson criterion, an ignited plasma is one where the fusion heating power is high enough to overcome all the physical processes that cool the fusion plasma, creating a positive thermodynamic feedback loop with rapidly increasing temperature. In inertially confined fusion, ignition is a state where the fusion plasma can begin "burn propagation" into surrounding cold fuel, enabling the possibility of high energy gain. While "scientific breakeven" (i.e., unity target gain) has not yet been achieved (here target gain is 0.72, 1.37 MJ of fusion for 1.92 MJ of laser energy), this Letter reports the first controlled fusion experiment, using laser indirect drive, on the National Ignition Facility to produce capsule gain (here 5.8) and reach ignition by nine different formulations of the Lawson criterion.

A scoping review of publicly available language tasks in clinical natural language processing

OBJECTIVE

To provide a scoping review of papers on clinical natural language processing (NLP) shared tasks that use publicly available electronic health record data from a cohort of patients.

MATERIALS AND METHODS

We searched 6 databases, including biomedical research and computer science literature databases. A round of title/abstract screening and full-text screening were conducted by 2 reviewers. Our method followed the PRISMA-ScR guidelines.

RESULTS

A total of 35 papers with 48 clinical NLP tasks met inclusion criteria between 2007 and 2021. We categorized the tasks by the type of NLP problems, including named entity recognition, summarization, and other NLP tasks. Some tasks were introduced as potential clinical decision support applications, such as substance abuse detection, and phenotyping. We summarized the tasks by publication venue and dataset type.

DISCUSSION

The breadth of clinical NLP tasks continues to grow as the field of NLP evolves with advancements in language systems. However, gaps exist with divergent interests between the general domain NLP community and the clinical informatics community for task motivation and design, and in generalizability of the data sources. We also identified issues in data preparation.

CONCLUSION

The existing clinical NLP tasks cover a wide range of topics and the field is expected to grow and attract more attention from both general domain NLP and clinical informatics community. We encourage future work to incorporate multidisciplinary collaboration, reporting transparency, and standardization in data preparation. We provide a listing of all the shared task papers and datasets from this review in a GitLab repository.

Hierarchical Annotation for Building A Suite of Clinical Natural Language Processing Tasks: Progress Note Understanding

Applying methods in natural language processing on electronic health records (EHR) data is a growing field. Existing corpus and annotation focus on modeling textual features and relation prediction. However, there is a paucity of annotated corpus built to model clinical diagnostic thinking, a process involving text understanding, domain knowledge abstraction and reasoning. This work introduces a hierarchical annotation schema with three stages to address clinical text understanding, clinical reasoning, and summarization. We created an annotated corpus based on an extensive collection of publicly available daily progress notes, a type of EHR documentation that is collected in time series in a problem-oriented format. The conventional format for a progress note follows a Subjective, Objective, Assessment and Plan heading (SOAP). We also define a new suite of tasks, Progress Note Understanding, with three tasks utilizing the three annotation stages. The novel suite of tasks was designed to train and evaluate future NLP models for clinical text understanding, clinical knowledge representation, inference, and summarization.

A Rare Case of COVID-19 Induced Multisystem Inflammatory Syndrome in Adult (MIS-A) Requiring Venoarterial Extracorporeal Membrane Oxygenation (VA-ECMO)

Introduction

MIS-A is a rare COVID-19 induced condition defined by fever, new-onset severe cardiac illness, rash, encephalopathy, and elevated inflammatory markers in the setting of positive serum COVID-19 antibodies. This inflammatory cascade can cause significant biventricular dysfunction and subsequent cardiogenic shock. Patients with MIS-A can require temporary cardiac support including VA-ECMO. We present a case of a patient requiring VA-ECMO secondary to MIS-A induced heart failure and cardiogenic shock, with eventual myocardial recovery.

Case Report

30-year-old male with type two diabetes was admitted with acute hypoxic respiratory failure, multiorgan failure, acute systolic biventricular heart failure, and COVID-19 infection four weeks prior. He was intubated and placed on vasopressors, antibiotics, and steroids for concerns for combined cardiogenic and septic shock. TTE noted global hypokinesis and 10-15% EF. EKG was sinus rhythm. He had mildly elevated troponins. Inflammatory markers including D-dimer, fibrinogen, and IL-6 were highly elevated. Despite antibiotics and supportive measures, the patient developed worsening hypoxia and hypotension. IVIG was also initiated, with deferral of plasmapheresis. At this time, MIS-A was suspected. The patient was approved for VA-ECMO as a means for bridging to cardiac recovery. He required VA-ECMO for four days, with ability to decannulate, extubate, and wean off vasopressors. COVID-19 antibody testing was positive. Infectious workup was negative, with the patient transitioned off antibiotics and steroid regimen after completing treatment course. Inflammatory markers improved. Repeat TTE noted 44% EF with improved biventricular function. Cardiac MRI one day later, noted 61% EF without evidence of scar, myocarditis, or perimyocarditis. He was discharged home after a total of 8 days of treatment with follow-ups scheduled.

Summary

This case highlights a severe presentation of MIS-A and showcases the benefit of VA-ECMO as a bridge to myocardial recovery. VA-ECMO has been shown to improve in-hospital survival and serve as a mechanism for cardiac recovery in acutely ill patients. Long-term cardiac effects and recovery rates post COVID-19 induced MIS-A remain unknown.

Low Anterior Cervical Approach Without Sternotomy or Clavicle Resection for Upper Thoracic Vertebra Corpectomy

The spine is the third most common site for metastatic disease following the lung and the liver. Approximately 60-70% of patients with metastatic cancer will have metastasis to the spine, but only 10% of these will be symptomatic. Metastases to the spine may involve the bone, epidural space, or the spinal cord. While chemotherapy and radiation therapy are the primary treatments for metastatic disease, spinal cord compression is an indication for surgical intervention. For vertebral body lesions, anterior vertebral reconstruction and stabilization also have the advantage of providing immediate stability to the vertebral column, but this anterior surgical approach to the upper thoracic spine is fraught with complications. The approach typically involves some combination of thoracotomy, sternotomy, or clavicle resection with anterior dissection into the superior mediastinum. To avoid unnecessary sternotomy and its associated complications, surgical access without sternotomy can be performed in certain cases. A sagittal MRI scan of the spine can be used to evaluate the level of the sternal notch in relation to the upper thoracic spine. If a tangential line can be drawn superior to the sternal notch and inferior to the level of the involved vertebra, surgical access without sternotomy can be performed. We present a case of a 52-year-old female with metastases to the upper thoracic vertebrae who underwent successful T2 corpectomy and T1-3 anterior fusion via a low anterior cervical approach, without sternotomy or clavicle resection.

A Review of Recent Work in Transfer Learning and Domain Adaptation for Natural Language Processing of Electronic Health Records

Objectives: We survey recent work in biomedical NLP on building more adaptable or generalizable models, with a focus on work dealing with electronic health record (EHR) texts, to better understand recent trends in this area and identify opportunities for future research.

Methods: We searched PubMed, the Institute of Electrical and Electronics Engineers (IEEE), the Association for Computational Linguistics (ACL) anthology, the Association for the Advancement of Artificial Intelligence (AAAI) proceedings, and Google Scholar for the years 2018-2020. We reviewed abstracts to identify the most relevant and impactful work, and manually extracted data points from each of these papers to characterize the types of methods and tasks that were studied, in which clinical domains, and current state-of-the-art results.

Results: The ubiquity of pre-trained transformers in clinical NLP research has contributed to an increase in domain adaptation and generalization-focused work that uses these models as the key component. Most recently, work has started to train biomedical transformers and to extend the fine-tuning process with additional domain adaptation techniques. We also highlight recent research in cross-lingual adaptation, as a special case of adaptation.

Conclusions: While pre-trained transformer models have led to some large performance improvements, general domain pre-training does not always transfer adequately to the clinical domain due to its highly specialized language. There is also much work to be done in showing that the gains obtained by pre-trained transformers are beneficial in real world use cases. The amount of work in domain adaptation and transfer learning is limited by dataset availability and creating datasets for new domains is challenging. The growing body of research in languages other than English is encouraging, and more collaboration between researchers across the language divide would likely accelerate progress in non-English clinical NLP.

Relationship between bone strength index of the hemiparetic tibial diaphysis and muscle strength in people with chronic stroke: influence of muscle contraction type and speed

This study was conducted to examine the association between the tibial bone strength index and leg muscle strength of different contraction types and speeds among people with chronic stroke. We found that concentric muscle power at moderate speed was more associated with tibial bone strength index than other types.

INTRODUCTION

To compare the influence of muscle strength of different contraction types and speeds on the bone strength index of tibial diaphysis in people with chronic stroke.

METHODS

Eighty individuals with chronic stroke (age: 62.6 ± 8.0 years; men/women: 46/34; post-stroke duration: 9.0 ± 5.4 years) underwent scanning of the tibia at the 66% site on both sides using peripheral quantitative computed tomography. Each participant was also evaluated for isometric and dynamic (at 60°/s and 120°/s) strength of knee flexors/extensors and ankle dorsiflexors/plantarflexors using an isokinetic dynamometer. For a given contraction type and speed, the strength values of the four muscle groups were summed to yield a composite score. Multiple regression analysis was used to identify the association of the percent side-to-side difference (%SSD) in tibial polar-stress-strain index (p-SSI) with %SSD in muscle strength of different contraction types and speeds.

RESULTS

The p-SSI and all muscle strength parameters on the paretic side had lower values than the non-paretic side (p ≤ 0.001). The %SSD in concentric muscle power at angular speed of 60°/s (R² = 0.317, p = 0.006) and 120°/s (R² = 0.298, p = 0.020) remained independently associated with that in p-SSI, after controlling for age, sex, body mass index, post-stroke duration, motor impairment, spasticity, and physical activity level. The effect of isometric strength and eccentric muscle power was not significant in multivariate analysis.

CONCLUSIONS

Concentric muscle power was more predictive of %SSD in p-SSI at the tibial diaphysis than other contraction types and may be an important target for intervention to promote bone health in people with chronic stroke.

Normalizing Clinical Document Titles to LOINC Document Ontology: an Initial Study

The normalization of clinical documents is essential for health information management with the enormous amount of clinical documentation generated each year. The LOINC Document Ontology (DO) is a universal clinical document standard in a hierarchical structure. The objective of this study is to investigate the feasibility and generalizability of LOINC DO by mapping from clinical note titles across five institutions to five DO axes. We first developed an annotation framework based on the definition of LOINC DO axes and manually mapped 4,000 titles. Then we introduced a pre-trained deep learning model named Bidirectional Encoder Representations from Transformers (BERT) to enable automatic mapping from titles to LOINC DO axes. The results showed that the BERT-based automatic mapping achieved improved performance compared with the baseline model. By analyzing both manual annotations and predicted results, ambiguities in LOINC DO axes definition were discussed.

Pre-training phenotyping classifiers

Recent transformer-based pre-trained language models have become a de facto standard for many text classification tasks. Nevertheless, their utility in the clinical domain, where classification is often performed at encounter or patient level, is still uncertain due to the limitation on the maximum length of input. In this work, we introduce a self-supervised method for pre-training that relies on a masked token objective and is free from the limitation on the maximum input length. We compare the proposed method with supervised pre-training that uses billing codes as a source of supervision. We evaluate the proposed method on one publicly-available and three in-house datasets using the standard evaluation metrics such as the area under the ROC curve and F1 score. We find that, surprisingly, even though self-supervised pre-training performs slightly worse than supervised, it still preserves most of the gains from pre-training.

Phase 1-2 Trial of Antisense Oligonucleotide Tofersen for SOD1 ALS

BACKGROUND

Tofersen is an antisense oligonucleotide that mediates the degradation of superoxide dismutase 1 (SOD1) messenger RNA to reduce SOD1 protein synthesis. Intrathecal administration of tofersen is being studied for the treatment of amyotrophic lateral sclerosis (ALS) due to SOD1 mutations.

METHODS

We conducted a phase 1-2 ascending-dose trial evaluating tofersen in adults with ALS due to SOD1 mutations. In each dose cohort (20, 40, 60, or 100 mg), participants were randomly assigned in a 3:1 ratio to receive five doses of tofersen or placebo, administered intrathecally for 12 weeks. The primary outcomes were safety and pharmacokinetics. The secondary outcome was the change from baseline in the cerebrospinal fluid (CSF) SOD1 concentration at day 85. Clinical function and vital capacity were measured.

RESULTS

A total of 50 participants underwent randomization and were included in the analyses; 48 participants received all five planned doses. Lumbar puncture-related adverse events were observed in most participants. Elevations in CSF white-cell count and protein were reported as adverse events in 4 and 5 participants, respectively, who received tofersen. Among participants who received tofersen, one died from pulmonary embolus on day 137, and one from respiratory failure on day 152; one participant in the placebo group died from respiratory failure on day 52. The difference at day 85 in the change from baseline in the CSF SOD1 concentration between the tofersen groups and the placebo group was 2 percentage points (95% confidence interval [CI], -18 to 27) for the 20-mg dose, -25 percentage points (95% CI, -40 to -5) for the 40-mg dose, -19 percentage points (95% CI, -35 to 2) for the 60-mg dose, and -33 percentage points (95% CI, -47 to -16) for the 100-mg dose.

CONCLUSIONS

In adults with ALS due to SOD1 mutations, CSF SOD1 concentrations decreased at the highest concentration of tofersen administered intrathecally over a period of 12 weeks. CSF pleocytosis occurred in some participants receiving tofersen. Lumbar puncture-related adverse events were observed in most participants. (Funded by Biogen; ClinicalTrials.gov number, NCT02623699; EudraCT number, 2015-004098-33.).

High-Brightness Continuous-Wave Electron Beams from Superconducting Radio-Frequency Photoemission Gun

Continuous-wave photoinjectors operating at high accelerating gradients promise to revolutionize many areas of science and applications. They can establish the basis for a new generation of monochromatic x-ray free electron lasers, high-brightness hadron beams, or a new generation of microchip production. In this Letter we report on the record-performing superconducting rf electron gun with CsK_{2}Sb photocathode. The gun is generating high charge electron bunches (up to 10  nC/bunch) and low transverse emittances, while operating for months with a single photocathode. This achievement opens a new era in generating high-power beams with a very high average brightness.

Experimental Demonstration of Hadron Beam Cooling Using Radio-Frequency Accelerated Electron Bunches

Cooling of beams of gold ions using electron bunches accelerated with radio-frequency systems was recently experimentally demonstrated in the Relativistic Heavy Ion Collider at Brookhaven National Laboratory. Such an approach is new and opens the possibility of using this technique at higher energies than possible with electrostatic acceleration of electron beams. The challenges of this approach include generation of electron beams suitable for cooling, delivery of electron bunches of the required quality to the cooling sections without degradation of beam angular divergence and energy spread, achieving the required small angles between electron and ion trajectories in the cooling sections, precise velocity matching between the two beams, high-current operation of the electron accelerator, as well as several physics effects related to bunched-beam cooling. Here we report on the first demonstration of cooling hadron beams using this new approach.

XAI-Explainable artificial intelligence

Explainability is essential for users to effectively understand, trust, and manage powerful artificial intelligence applications.

Toward a clinical text encoder: pretraining for clinical natural language processing with applications to substance misuse

OBJECTIVE

Our objective is to develop algorithms for encoding clinical text into representations that can be used for a variety of phenotyping tasks.

MATERIALS AND METHODS

Obtaining large datasets to take advantage of highly expressive deep learning methods is difficult in clinical natural language processing (NLP). We address this difficulty by pretraining a clinical text encoder on billing code data, which is typically available in abundance. We explore several neural encoder architectures and deploy the text representations obtained from these encoders in the context of clinical text classification tasks. While our ultimate goal is learning a universal clinical text encoder, we also experiment with training a phenotype-specific encoder. A universal encoder would be more practical, but a phenotype-specific encoder could perform better for a specific task.

RESULTS

We successfully train several clinical text encoders, establish a new state-of-the-art on comorbidity data, and observe good performance gains on substance misuse data.

DISCUSSION

We find that pretraining using billing codes is a promising research direction. The representations generated by this type of pretraining have universal properties, as they are highly beneficial for many phenotyping tasks. Phenotype-specific pretraining is a viable route for trading the generality of the pretrained encoder for better performance on a specific phenotyping task.

CONCLUSIONS

We successfully applied our approach to many phenotyping tasks. We conclude by discussing potential limitations of our approach.